TEST #954 RESULTS
05/29/2025 - 9:11 PM
ACCELERATOR
MODEL
46.2
tokens/s
generation
723
ms
time to first token
1816
tokens/s
prompt
488
LocalScore
HOW YOU STACK UP
Explore All ResultsQwen2.5 14B Instruct - Q4_K - Medium
SYSTEM
CPU
Intel Xeon Gold 6338N CPU @ 2.20GHz (icelake-server)
RAM
503.4GB
OS
Linux
Kernel Release
5.15.0-139-generic
Architecture
x86_64
Version
Cosmopolitan 3.9.7 MODE=x86_64; #149~20.04.1-Ubuntu SMP Wed Apr 16 08:29:56 UTC 2025
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
2170
tokens/s
47.5
tokens/s
493
ms
pp4096+tg256
1693
tokens/s
42.0
tokens/s
2.44
sec
pp2048+tg256
1980
tokens/s
45.6
tokens/s
1.06
sec
pp2048+tg768
1984
tokens/s
45.1
tokens/s
1.05
sec
pp1024+tg1024
2158
tokens/s
46.7
tokens/s
496
ms
pp1280+tg3072
2091
tokens/s
44.4
tokens/s
633
ms
pp384+tg1152
2228
tokens/s
47.7
tokens/s
209
ms
pp64+tg1024
1618
tokens/s
48.4
tokens/s
60
ms
pp16+tg1536
423
tokens/s
48.0
tokens/s
58
ms